摘要
针对动态定位过程中摄像机的运动使得目标成像模糊 ,提出了一种基于模糊图像矩进行三维动态定位的控制系统 .该系统依靠单摄像机在笛卡尔坐标系中的三维平动迅速接近环境中的静止目标 ,且与静止目标的三维距离符合给定值 .证明了图像的零阶和一阶矩是模糊不变量 ,且包含目标的三维信息 .推导出被控对象的状态方程 ,给出了系统的控制结构及控制律 .仿真结果表明 。
In the process of localization, images of an object is often blurred for the moving of camera. To solve this problem, a dynamic localization system based on moments of blurred images was proposed. Using this system, a camera can be driven to quickly approach a static object in 3D translatory velocity and keep required 3D distance to the object. It was proved in the paper that 0 and 1 order image moments not only are invariant of a blurred image, but also include 3D information of an object. The state equations of controlled system, control structure, and a control algorithm were also deduced. At last, some simulating results of the control were given to prove this kind of dynamic localization system can acquire a satisfying accuracy.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2003年第6期951-955,共5页
Journal of Shanghai Jiaotong University
关键词
定位
模糊图像
图像矩
雅可比矩阵
localization
blurred image
image moment
Jacobian matrices
